The proposed work is to take quantitative approach to study the effect of interplay between chimeric antigen receptor (CAR) affinity and antigen density on CAR T cell efficacy and toxicity. Aside from a limited number of surface-expressed tumor neoantigens, there is a widely acknowledged paucity of tumor-specific targets. Selectively tuning a CAR?s affinity for its target molecule can maintain its ability to lyse tumor cells exhibiting high density target expression while preventing this occurrence in normal cells with lower, basal expression. However, our current understanding of how the threshold for T cell activation and cytotoxicity is defined by the interplay between CAR affinity and antigen density is still in its infancy. Relevant studies to date have examined this relationship using CARs with crude affinity variations and target cells devoid of quantitative, antigen density measurements. Furthermore, the potential for systemic off-tumor toxicity mediated by affinity- tuned CAR T cells has yet to be evaluated in a pre-clinical environment where the CAR in question can cross- react with natural levels of its cognate host antigen, thereby rigorously performing risk assessment and therapeutic index studies of off-tumor CAR reactivity. To systematically examine the affinity tuned CAR paradigm, we have designed a CAR that specifically targets intercellular adhesion molecule (ICAM)-1, a molecule that is over-expressed in multiple tumors but which retains weak expression in healthy tissue. This CAR?s antigen recognition domain is derived from the inserted (I) domain of an integrin called lymphocyte function-associated antigen (LFA)-1, which we previously adapted by introducing point mutations such that step-wise variations of ICAM-1 affinity could be derived ranging from 1 mM to 1 nM Kd or 106-fold. Human LFA-1 I domain binds murine and human ICAM-1 with comparable affinity, allowing simultaneous in vivo evaluation of CAR reactivity against human ICAM-1 expressing, transplanted tumor tissue alongside any potential on-target, off-tumor reactivity/toxicity against murine ICAM-1. We have also recently developed and evaluated a genetic reporter, human somatostatin receptor 2 (SSTR2), that enables quantitative, non-invasive real-time monitoring and tracking of T cells using positron emission tomography-computed tomography (PET/CT) and clinically approved imaging reagents. In this proposal, we will 1) determine threshold CAR affinity that restricts CAR T cell cytotoxicity to tumors with over-expressed antigens, and 2) examine the influence of CAR affinity and antigen density on spatiotemporal kinetics of T cell expansion and contraction, and on the rate of tumor elimination, relapse, and systemic toxicity. The outcome of this study has the potential to significantly impact the study of CAR T cells by providing a quantitative framework for CAR T cell design and evaluation to maximize their therapeutic index.